Integrative Kinome Profiling Identifies Mtorc1/2 Inhibition As Treatment Strategy in Ovarian Clear Cell Carcinoma

Published OnlineFirst April 23, 2018; DOI: 10.1158/1078-0432.CCR-17-3060

Cancer Therapy: Preclinical Clinical Cancer Research Integrative Kinome Proﬁling Identiﬁes mTORC1/2 Inhibition as Treatment Strategy in Ovarian Clear Cell Carcinoma Joseph J. Caumanns1, Katrien Berns2, G. Bea A. Wisman1, Rudolf S.N. Fehrmann3, Tushar Tomar1, Harry Klip1, Gert J. Meersma1, E. Marielle Hijmans2, Annemiek M.C. Gennissen2, Evelien W. Duiker4, Desiree Weening5, Hiroaki Itamochi6, Roelof J.C. Kluin2, Anna K.L. Reyners3, Michael J. Birrer7, Helga B. Salvesen8, Ignace Vergote9, Els van Nieuwenhuysen9, James Brenton10, E. Ioana Braicu11, Jolanta Kupryjanczyk12, Beata Spiewankiewicz13, Lorenza Mittempergher2, Rene Bernards2, Ate G.J. van der Zee1, and Steven de Jong3

Abstract

Purpose: Advanced-stage ovarian clear cell carcinoma Results: We identiﬁed several putative driver mutations in (OCCC) is unresponsive to conventional platinum-based che- kinases at low frequency that were not previously annotated in motherapy. Frequent alterations in OCCC include deleterious OCCC. Combining mutations and copy-number alterations, 91% mutations in the tumor suppressor ARID1A and activating of all tumors are affected in the PI3K/AKT/mTOR pathway, mutations in the PI3K subunit PIK3CA.Inthisstudy,weaimed the MAPK pathway, or the ERBB family of receptor tyrosine to identify currently unknown mutated kinases in patients with kinases, and 82% in the DNA repair pathway. Strong p-S6 staining OCCC and test druggability of downstream affected pathways in patients with OCCC suggests high mTORC1/2 activity. We in OCCC models. consistently found that the majority of OCCC cell lines are Experimental Design: In a large set of patients with OCCC (n ¼ especially sensitive to mTORC1/2 inhibition by AZD8055 and 124), the human kinome (518 kinases) and additional cancer- not toward drugs targeting ERBB family of receptor tyrosine related genes were sequenced, and copy-number alterations were kinases or DNA repair signaling. We subsequently demonstrated determined. Genetically characterized OCCC cell lines (n ¼ 17) the efﬁcacy of mTORC1/2 inhibition in all our unique OCCC and OCCC patient–derived xenografts (n ¼ 3) were used for drug patient–derived xenograft models. testing of ERBB tyrosine kinase inhibitors erlotinib and lapatinib, Conclusions: These results propose mTORC1/2 inhibition as the PARP inhibitor olaparib, and the mTORC1/2 inhibitor an effective treatment strategy in OCCC. Clin Cancer Res; 1–13. 2018 AZD8055. AACR.

Introduction ian cancer. The majority of patients with OCCC are diagnosed at In the United States, ovarian cancer is the ﬁfth-leading cause of an early stage (57%–81% at stage I/II) and have better overall cancer deaths in women (1). Ovarian clear cell carcinoma survival compared with stage-matched high-grade serous (HGS) (OCCC) is the second most common subtype of epithelial ovar- ovarian cancer, the most common subtype of ovarian cancer. In

1Department of Gynecologic Oncology, Cancer Research Center Groningen, Medical University, Berlin, Germany. 12Department of Pathology and Laboratory University Medical Center Groningen, University of Groningen, Groningen, the Diagnostics, Maria Sklodowska-Curie Memorial Cancer Center and Institute of Netherlands. 2Division of Molecular Carcinogenesis, The Netherlands Cancer Oncology, Warsaw, Poland. 13Department of Gynecologic Oncology, Maria Institute, Amsterdam, the Netherlands. 3Department of Medical Oncology, Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Warsaw, Cancer Research Center Groningen, University Medical Center Groningen, Poland. University of Groningen, Groningen, the Netherlands. 4Department of Pathology and Medical Biology, Cancer Research Center Groningen, University Medical Note: Supplementary data for this article are available at Clinical Cancer Center Groningen, University of Groningen, Groningen, the Netherlands. Research Online (http://clincancerres.aacrjournals.org/). 5Department of Genetics, Cancer Research Center Groningen, University Med- ical Center Groningen, University of Groningen, Groningen, the Netherlands. In memory of H. Klip and H.B. Salvesen. 6Department of Obstetrics and Gynecology, Iwate Medical University School of Corresponding Author: Steven de Jong, Cancer Research Center Groningen, Medicine, Morioka, Iwate, Japan. 7Center for Cancer Research, The Gillette University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 Center for Gynecologic Oncology, Massachusetts General Hospital, Harvard GZ Groningen, the Netherlands. Phone: 31-50-3612964; Fax: 31-50-3614862; Medical School, Boston, Massachusetts. 8Department of Obstetrics and Gyne- E-mail: [email protected] cology, Haukeland University Hospital, Bergen, Norway. 9Department of Gynae- cology and Obstetrics, Leuven Cancer Institute, University Hospitals Leuven, doi: 10.1158/1078-0432.CCR-17-3060 Leuven, Belgium. 10Cancer Research UK Cambridge Research Institute, Li Ka Shing Centre, Cambridge, United Kingdom. 11Department of Gynecology, Charite 2018 American Association for Cancer Research.

www.aacrjournals.org OF1

Caumanns et al.

research in OCCC cell lines demonstrated sensitivity to PI3K/ Translational Relevance mTOR dual inhibitors and AKT inhibitors, although PIK3CA Advanced-stage ovarian clear cell carcinoma (OCCC) is less mutations did not predict sensitivity to these inhibitors (16, 17). responsive to platinum-based chemotherapy compared with In the present study, we aimed to identify novel targetable high-grade serous ovarian carcinoma. Our in-depth analyses mutations by means of high-coverage sequencing of all protein of a large set of patients with OCCC reveal numerous genomic kinase genes, referred to as the kinome, and of a subgroup of alterations related to activation of mTORC1/2. High sensitiv- cancer-related genes in a large set of OCCC. In addition, we ity, especially to inhibitors targeting both mTORC1 and determined copy-number gains and losses in kinases and other mTORC2, is observed in a large OCCC cell line panel. Similar genes of OCCC tumors using high-coverage single-nucleotide results are obtained in OCCC patient–derived xenografts, polymorphism (SNP) arrays. To detect kinase mutations and signifying the clinical implications of these genomic altera- CNA at both high and low frequencies, we used a large cohort tions. Targeting of mTORC1 in combination with standard of 124 untreated primary OCCC tumors and most of the available chemotherapy did not improve overall survival in patients OCCC cell lines (n ¼ 17). Finally, we functionally validated with OCCC. Therefore, mTORC1/2 inhibitors, currently eval- several candidate targets in OCCC cell lines and unique OCCC uated in phase II clinical trials, are proposed for the treatment patient-derived xenograft (PDX) models. Our results indicate of OCCC. Based on the mutational landscape in OCCC, our in mTORC1/2 inhibition as an approach to guide future develop- vitro results, and the toxicity observed with dual inhibitors of ment of therapeutic strategies for OCCC. PI3K and mTORC1/2, future treatment combinations of mTORC1/2 inhibitors with either PI3K or MEK inhibitors Materials and Methods could be considered to improve clinical benefit for patients Sample collection with OCCC. Primary tumor samples from 124 patients with OCCC and 47 paired control blood samples were prospectively collected from Belgium, Germany, Norway, Poland, the Netherlands, the UK, and the United States. All patients gave written informed consent for samples to be collected, and the corresponding ethical review contrast, patients with OCCC diagnosed at late stage respond boards approved the study. Tumor samples had to contain 40% poorly to standard platinum-based chemotherapy compared with tumor cells, of which 70% was OCCC, as determined by patients with late-stage HGS ovarian carcinoma (2). In recent experienced gynecologic oncology pathologists. We obtained years, genetic studies in relatively small patient groups have 17 human OCCC cell lines: TOV21G (ATCC); RMG1, RMG2, revealed the mutational landscape in OCCC. The SWI-SNF chro- OVMANA, OVTOKO, and HAC2 (JCRB Cell Bank); JHOC5 matin remodeling complex DNA-binding AT-rich interactive (RIKEN Cell Bank); OVCA429 (Cell Biolabs); OVSAYO, TUOC1, domain 1A gene (ARID1A) has been shown to be deleteriously KK, OVAS, SMOV2, and KOC7C (Dr. Hiroaki Itamochi, mutated in 40% to 57% of patients with OCCC, the highest Tottori University School of Medicine, Tottori, Japan); ES2 percentage found in any cancer (3, 4). Loss of ARID1A protein, (Dr. Els Berns, Erasmus MC, Rotterdam, the Netherlands); TAYA being a key component of the complex, may affect the expression (Dr. Yasushi Saga, Jichi Medical University, Yakushiji, of many genes (5). Activation of the PI3K/AKT/mTOR pathway, Shimotsuke-shi, Tochigi, Japan); and OV207 (Dr. Vijayalakshmi implicated in survival, protein synthesis, and proliferation, is Shridhar, Mayo Clinic). All cell lines were maintained in RPMI another major player in OCCC. PIK3CA, encoding the catalytic supplemented with 10% FCS, 100 mg/mL Penicillin/Streptomy- domain of PI3K, contains activating mutations in 30% to 40% of cin and 2 mM L-glutamine. All cell lines were tested by short patients with OCCC, whereas expression of the PI3K antagonist tandem repeat profiling and were mycoplasma free. All cell PTEN is diminished in 40% of patients with OCCC (4, 6, 7). lines were kept in culture for a maximum of 50 passages. Furthermore, mutations in the oncogene KRAS and the tumor- suppressor gene TP53 have been identified in 4.7% to 14% and 10% to 15% of patients with OCCC, respectively (1, 4, 6, 8, 9). Kinome sequencing In addition to mutational aberrations, copy-number alterations Library construction, exome capture, and sequencing. From 124 (CNAs) have been found in OCCC tumor samples in the proto primary fresh-frozen OCCC tumors and 47 paired controls, 3-mg oncogene ZNF217, tumor-suppressor genes, cyclin-dependent DNA was prepared for sequencing using the following steps. kinase inhibitors CDKN2A and CDKN2B, and the membrane Genomic DNA was sheared to produce 300 bp fragments (Covaris receptor oncogene MET (10–12). S220); using SureSelect Target enrichment & Human Kinome Kit The identification of the most frequently mutated genes (Agilent technologies), kinase exons were tagged and captured; ARID1A and PIK3CA may lead to new therapeutic strategies. In using biotinylated RNA library baits and streptavidin beads, exons particular, the effects of ARID1A loss are being investigated, and were amplified and loaded on a HiSeq2500 Illumina sequencer vulnerabilities in ARID1A-mutant cancers are being identified. using paired-end sequencing according to the manufacturer's Synthetic lethal interactions have recently been demonstrated in protocols. The SureSelect Human Kinome Kit captures exons from ARID1A-mutant OCCC cancer cell lines by shRNA-mediated 518 kinases, 13 diglyceride kinases, 18 PI3K domain and regu- suppression of ARID1B, a homolog of ARID1A, as well as chem- latory component genes, and 48 cancer-related genes (Supple- ical inhibition of the histone H3 methyltransferase EZH2 and mentary Table S1). After sequencing, raw data were mapped to the histone deacetylase HDAC6 (13–15). PI3K signaling–mediated human reference sequence NCBI build 37 (hg19) and processed tumor addiction through the well-studied hypermorphic mutant according to our sequencing pipeline (Supplementary Fig. S1). forms of PIK3CA (E545 and H1047) was studied extensively in Genome Analysis Toolkit (GATK, version 1.0.5069) was used for multiple cancer types including OCCC. Recent translational indel realignment and base quality recalibration on BAM files. See

OF2 Clin Cancer Res; 2018 Clinical Cancer Research

Kinome Proﬁling Identiﬁes mTORC1/2 Sensitivity in OCCC

the Supplementary Methods for further details on kinome (TP53, ATM, and PRKDC), ERBB family of receptor tyrosine sequencing. kinase genes (ERBB3 and EGFR), and chromatin remodeling genes (ARID1A). The frequencies of previously described mutated SNP array genes observed in our study were in agreement with such fre- SNP genotyping and quality control. Genome-wide SNP genotyp- quencies reported earlier in smaller studies of OCCC (Fig. 1A; ing was performed with HumanOmniExpressExome-8BeadChip refs. 3, 4, 6, 8, 9). (Illumina) containing >900K SNPs, including >273K function- ARID1A mutations were analyzed using Haloplex sequencing al exomic markers to determine CNA in 108 primary OCCC only. Of 54 ARID1A-mutant tumors, 18 tumors contained homo- tumors and 17 OCCC cell lines. DNA sample processing, zygous frameshift or stop-gain mutations, 13 tumors harbored hybridization, labeling, scanning, and data extraction were more than one heterozygous mutation, whereas 23 tumors con- performed according to Illumina infinium 2 protocol. Illumina tained a single heterozygous frameshift or stop-gain mutation GenomeStudio software was used for primary sample assess- (Fig. 1B). The identified proportion of each type of mutation in ment and SNP call rate quality control of SNP intensity output ARID1A matches those reported in earlier studies (3, 4). ARID1A- files. See the Supplementary Methods for further details on SNP mutant and wild-type tumors did not show differential mutation array analysis. incidence in kinome genes (7.4 vs. 6.6 mutations per tumor on average). fi In vitro inhibitor screening Statistical binomial univariate testing of all 851 identi ed fi < The 17 OCCC cell lines were plated in 384-well plates with 250 kinome mutations revealed 11 signi cantly mutated genes (P to 2,000 cells/well using an automatic cell dispenser. Cell lines 0.05) relative to background mutations. These predicted onco- were incubated with varying concentrations of AZD8055 (Axon genic drivers were PIK3CA, KRAS, TP53, PTEN, AKT1, PIK3R1, Medchem), gefitinib (Selleckchem), lapatinib (Selleckchem), or FBXW7, ERBB3, ATM, CHEK2, and MYO3A (Supplementary olaparib (Axon Medchem) for 6 days. Cell viability was subse- Table S3). PIK3CA and KRAS exhibited mutations in established quently monitored using CellTiter-Blue cell viability assay (Pro- hotspot sites in aa Q542, Q545, and H1047 in PIK3CA and G12 in mega). Due to large-cell density-dependent variability in drug KRAS, whereas TP53 mutations were distributed across the TP53 – sensitivity, for each cell line, the lowest and highest IC were DNA-binding domain (Fig. 1C E). 50 fi removed. The cell lines CAPAN1, OE19, LOVO, OVCAR3, Interestingly, we identi ed mutations in three genes not pre- OVCAR4, and SKOV3 were exposed to AZD8055 to serve as viously described in OCCC (AKT1, PIK3R1, and ERBB3) and with sensitive and resistant controls. an established role in PI3K/AKT/mTOR pathway activation. AKT1 missense mutations were identified in six tumors (4.9%). AKT1 is one of the key components of the PI3K/AKT/mTOR cascade. AZD8055 in vivo administration The PH domain of AKT1 interacts with its kinase domain and All animal experiments were approved by the Institutional maintains the protein in a closed and inactive state (18). One Animal Care and Use Committee of the University of Groningen candidate somatic mutation, D323N, was located in the kinase (Groningen, the Netherlands) and carried out in accordance with domain. The mutations R25H, L52R (occurring in three tumors), the approved guideline "code of practice: animal experiments in and W80R (both described in COSMIC) were all located in the cancer research" (Netherlands Inspectorate for Health Protection, AKT1 PH domain (Supplementary Fig. S2A). PIK3R1 was mutated Commodities and Veterinary Public Health, 1999). See the Sup- in nine tumors (7.3%), of which seven tumors carried mutations plementary Methods for further details on AZD8055 in vivo in the inter-SH2-1–SH2-2 domain of the protein (aa 429–623). administration. This domain binds PIK3CA and is required for the inhibitory role of PIK3R1 on PIK3CA (19). Two of the inter-SH2 domain Additional methods mutations are described in COSMIC: E439 and T576 in-frame K-means consensus clustering of kinome mutations and CNA, deletions. In addition, we identified six novel inter-SH2 domain mRNA level determination, flow cytometry, MTT assays, long- mutations; two somatic missense (aa N453D and T471S, term proliferation assays, Western blotting, immunohistochem- together in one tumor) and two somatic in-frame deletions ical analysis, data visualization, and statistical analysis are (aa QF455 and RE461 ), one candidate somatic frameshift described in the Supplementary Methods. (starting in aa 582), and one candidate somatic stop-gain mutation (aa 571; Supplementary Fig. S2B). In ERBB3,we Results identified 10 missense mutations across eight distinct tumors Kinome sequencing analysis (5.7%). At the cell membrane, ERBB3 can dimerize with other Across all OCCC tumors (n ¼ 122) and OCCC cell lines (n ¼ ERBB family members and regulate downstream kinase signal- 17), 95.9% of all bases had >20 read coverage for variant calling. ing. Six mutations were located across the extracellular domains The mean coverage depth for aligned reads was 99.3x. On average, of ERBB3, and two could be identified as somatic. The D297Y 1.17% of the sequenced genes were mutated per patient. Genes mutation has been described as a hotspot location in ovarian with a high mutation frequency (>4%) across all OCCC tumors and colorectal cancers (20). Furthermore, three mutations were are shown in Fig. 1A. Resequencing of these 19 genes using in the intracellular C-terminal domain and one in the kinase Haloplex confirmed 227 of the 234 mutations (97%) that were domain (Supplementary Fig. S2C). Of the other significantly originally identified by kinome sequencing using the same tumor mutated genes, F-Box and WD Repeat Domain Containing 7 DNA. The majority of genes with a high mutation frequency are (FBXW7) has been assigned a role in PI3K regulation and DNA implicated in well-known cancer-related pathways like the PI3K/ repair. These mutations are mainly found in its WD repeats AKT/mTOR pathway (PIK3CA, PTEN, PIK3R1, and AKT1), MAPK (21). ATM Serine/Threonine Kinase (ATM) and Checkpoint signaling transduction pathway (KRAS), DNA repair pathway kinase 2 (CHEK2) are designated as DNA repair genes, whereas

www.aacrjournals.org Clin Cancer Res; 2018 OF3

Caumanns et al.

Figure 1. Most frequent OCCC mutations. A, Frequently mutated genes in OCCC as identiﬁed by kinome sequencing in 122 OCCC tumors using a 4% cutoff. ARID1A mutations (n ¼ 54 tumors) were revealed using Haloplex sequencing on 116 OCCC tumors. Mutated genes involved in PI3K (PIK3CA, PIK3R1, PTEN, AKT1, and FBXW7; n ¼ 54, n ¼ 9, n ¼ 8, n ¼ 6, and n ¼ 5 tumors, respectively), MAPK (KRAS, n ¼ 19 tumors), or DNA repair signaling (TP53, ATM, PRKDC, and FBXW7; n ¼ 14, n ¼ 11, n ¼ 10, and n ¼ 5 tumors, respectively) and ERBB family of receptor tyrosine kinases (ERBB3 and EGFR, n ¼ 8andn ¼ 5 tumors) are among the frequently mutated genes. Schematics of identiﬁed mutations in known OCCC-mutated genes (B) ARID1A,(C) PIK3CA,(D) KRAS, and (E) TP53. Mutation marks are shown in black (truncating), red (SIFT and PolyPhen damaging prediction), yellow (SIFT or PolyPhen damaging prediction), or white (SIFT and PolyPhen benign prediction). Mutation effects are indicated with a black spot when paired control was available and written in black (previously described mutation) or red (novel mutations).

OF4 Clin Cancer Res; 2018 Clinical Cancer Research

Kinome Proﬁling Identiﬁes mTORC1/2 Sensitivity in OCCC

Myosin IIIA (MYO3A) is an actin-dependent motor protein. and ARID1A or PIK3CA mutations was observed in a few cases. In Mutations in FBXW7 and ATM were previously described in total 108 of 122 tumors (89%) comprised one or more mutations OCCC (Supplementary Fig. S2D–S2G; ref. 22). in genes belonging to the PI3K/AKT/mTOR pathway, MAPK pathway, DNA repair pathway, and ERBB family of receptor Pathway analysis of kinome mutations tyrosine kinases (Supplementary Fig. S3A). We also investigated the pathways in which genes with a high Evidently, there is a large overlap in tumors that were both mutation frequency are involved. To this end, all low-mutation ARID1A mutant and harbored a mutation in DNA repair frequency genes (<4%) present in PI3K/AKT/mTOR, MAPK, and genes or PI3K/AKT/mTOR, MAPK, and ERBB family of receptor DNA repair pathways, the ERBB family of receptor tyrosine tyrosine kinase genes (Supplementary Fig. S3B). Subdividing kinases and mutations in ARID1A were mapped per tumor in tumors into ARID1A heterozygous-mutated and homozygous- order to determine mutational spectra across these signaling mutated tumors did not change the overlap in mutations in PI3K/ pathways in OCCC. One or more mutations were found in AKT/mTOR, MAPK, and ERBB family of receptor tyrosine PI3K/AKT/mTOR pathway genes in 76 patients (62.3%), in MAPK kinase genes or DNA repair genesinbothgroups(datanot pathway genes in 22 patients (18%), in the ERBB family of shown). ARID1A wild-type tumors more frequently contained receptor tyrosine kinases in 18 patients (14.8%), and in DNA mutationsinDNArepairgenesinadditiontomutationsin repair pathway genes in 46 patients (37.7%). The co-occurrence of PI3K/AKT/mTOR, MAPK, or the ERBB family of receptor tyro- mutations in both ARID1A and PIK3CA in this large set of OCCC sine kinase genes as compared with ARID1A-mutant tumors tumors was in accordance with the findings from a previous, (49% vs. 26%, P ¼ 0.0385). smaller study (Supplementary Fig. S3A; refs. 4, 23). PIK3R1 mutations were mutually exclusive with PIK3CA-mutant tumors CNA analysis (P ¼ 0.043), and a trend was observed for PTEN with PIK3CA- In 108 SNP-genotyped primary OCCC tumors, GISTIC (Geno- mutant tumors (P ¼ 0.0756). In general, TP53-mutant tumors mic Identification of Significant Targets in Cancer) analysis (including pure OCCCs) were mutually exclusive with both annotated 324 significantly amplified genes located in 48 focal ARID1A- and PIK3CA-mutant tumors (P ¼ 0.0031), in agreement regions. We identified amplification of ZNF217 (20q13.20), a with previous literature (5). Surprisingly, co-occurrence of TP53 transcriptional regulator previously described in OCCC (11), in

Figure 2. Kinase CNA in OCCC. Significant CNA in kinases across 108 OCCC tumors as determined by GISTIC analysis. All kinases and cancer-related genes from the kinome sequencing gene panel that were focally significantly amplified (red) or deleted (blue) are vertically indicated along the chromosomes. Chromosomal location and total amount of tumors (between brackets) harboring the event are annotated with each gene name. The false discovery rate (FDR) 0.05 threshold, indicated by the green line, and G-score are shown along the horizontal axis.

www.aacrjournals.org Clin Cancer Res; 2018 OF5

Caumanns et al.

29 tumors (27%, P ¼ 0.014). Using GISTIC analysis, we subse- 3B), whereas the DNA repair pathway was affected in 82% of quently found multiple kinases and other cancer-related genes all tumors (Fig. 3C). among the significantly amplified genes (Fig. 2). One recurrently amplified region (3q26.2) contained multiple Kinome profile reveals tumor clusters with differential survival kinases: PIK3CA, EPHB3, MAPK3K13, PAK2, PRKCI, TNIK, TNK2, Clinical data were available for a subset of patients (n ¼ 70; and DGKG (14 tumors, 13%, P ¼ 0.046). The cancer-related gene Supplementary Table S4). Disease-specific survival analysis TERT (5p15.33), included in kinome sequencing, was amplified revealed that ARID1A, PIK3CA,orARID1A plus PIK3CA altera- in 12 tumors (11%, P ¼ 0.044). EGFR (7p11.2) emerged as the tions were not related to survival (Supplementary Fig. S5A– most significant recurrently amplified kinase; it was present in 18 S5C). Kinome mutations and CNA events of all 106 tumors OCCC tumors (17%, P < 0.0001). Other recurrently amplified were integrated, and the tumors were grouped using K-means kinases included ERBB2 (17q12), which was amplified in 21 consensus clustering. Maximum cluster number was set at 8, tumors (19%, P ¼ 0.002), and the chromatin-associated and because more clusters only marginally decreased friction (Sup- transcriptional control-related kinase TRIM28 (19q13.43), which plementary Fig. S5D–S5F). Most tumors grouped together in was amplified in 52 tumors (48%, P ¼ 0.0035). cluster 1 (n ¼ 53), 3 (n ¼ 27), or 5 (n ¼ 13). A trend for worse Furthermore, a total of 118 significantly deleted genes were disease-specific survival was demonstrated with cluster 3 com- identified in 62 focally deleted regions. Two significantly deleted pared with all other clusters (P ¼ 0.0638; Supplementary kinases could be identified: SIK3 (11q23.3) and the transcrip- Fig. S5G and S5H), which became highly significant upon tional repressor CDK8 (13q12.13; Fig. 2). In addition, 3 known selection for patients with advanced-stage OCCC (P < 0.001, cancer-related genes included in kinome sequencing were signif- n ¼ 31; Supplementary Fig. S5I). Grouping of mutation and icantly deleted. The cell-cycle regulators CDKN2A and CDKN2B CNA status for each cluster did not reveal unique genes in (both located on 9p21.3) were deleted in 16 tumors (15%, cluster 3 (Supplementary Fig. S5J and S5K). P ¼ 0.024), and PALB2 (16p12.2), an essential chaperone of BRCA2, was deleted in 33 tumors (31%, P < 0.0001). Kinome profile–based inhibitor screen identifies mTORC1/2 A separate GISTIC analysis was implemented to compare inhibition susceptibility ARID1A-mutant tumors (n ¼ 45) with ARID1A wild-type (n ¼ Kinome sequencing and SNP data of 17 OCCC cell lines 63) tumors (Supplementary Fig. S4A). TRIM28 amplification and presented similar alteration frequencies in PI3K/AKT/mTOR, CDK8 deletion were significantly retained only in ARID1A- MAPK, and DNA repair pathway genes and the ERBB family mutant tumors (P < 0.0001 and P ¼ 0.0067, respectively), whereas of receptor tyrosine kinases as those observed in patients with EGFR amplification was significantly retained only in ARID1A OCCC (Fig. 4A and B). However, an overrepresentation of wild-type tumors (P ¼ 0.0054; Supplementary Fig. S4B and S4C). ARID1A and TP53 mutations and an underrepresentation of KRAS mutations were found. Due to the resemblance of OCCC Integration of kinome mutations and CNA patient alterations with those in the cell line panel, we decided Both kinome sequencing and SNP data were available for 106 to screen for kinase inhibition vulnerabilities of the PI3K/AKT/ tumors. After merging all mutations and CNA events in the PI3K/ mTOR and MAPK pathway downstream targets mTORC1/2 and AKT/mTOR, MAPK, and DNA repair pathway and ERBB family of the ERBB receptor tyrosine kinases EGFR and ERBB2. In addi- receptor tyrosine kinases, we identified at least one event in 103 of tion, the DNA repair pathway was targeted. High-throughput 106 tumors analyzed (97%; Fig. 3A). drug sensitivity testing revealed nanomolar range efficacy Amplification incidence of >10% was found in the PI3K/AKT/ against the mTORC1/2 inhibitor AZD8055 in all 17 OCCC mTOR-related genes AKT2 (17%), PIK3R2 (14.2%), PIK3CA cell lines tested (Fig. 5A). Sensitivity appeared to be irrespective (12.3%), AKT1 (10.4%), the ERBB family of receptor tyrosine of PIK3CA or ARID1A mutation status, suggesting that altera- kinases ERBB2 (19.8%) and EGFR (17%), and the DNA repair tions upstream of mTORC1/2 may explain the comprehensive genes PRKDC (24%) and CHEK2 (12.3%). AKT2, PIK3R2, ERBB2, AZD8055 susceptibility. COSMICs cancer cell line drug screen- EGFR, and CHEK2 primarily contained amplifications, from ingdata(n ¼ 913 cell lines) with AZD8055 (Cancerrxgene.org) which only ERBB2 (n ¼ 2) and EGFR (n ¼ 1) presented tumors and temsirolimus (targeting mTORC1) indicate that OCCC cell that were both amplified and mutated. PIK3CA was mostly lines are among the most sensitive ovarian cancer cell lines (Fig. mutated, yet 4 PIK3CA-mutant tumors carried both an amplifi- 5B and Supplementary Fig. S6A). cation and a mutation. Amplifications and mutations never co- MTT assay–based IC50 determination in 14 of 17 OCCC cell occurred in AKT1 and PRKDC. lines demonstrated higher susceptibility to AZD8055 and the Deletion incidence of >10% was observed in the PI3K/AKT/ mTORC1 inhibitor everolimus in comparison with 3 HGS ovar- mTOR-related gene PIK3C3 (18.9%) and the DNA repair genes ian cancer cell lines (Fig. 5C). In contrast to AZD8055, prolifer- PALB2 (30.2%), ATM (17.9%), and BRCA2 (17%). PIK3C3, ation inhibition by everolimus displayed a plateau phase in a PALB2, and BRCA2 primarily harbored deletions, whereas ATM wide concentration range and a large variation in sensitivity deletions and mutations co-occurred in three tumors. among OCCC cell lines (Fig. 5D and Supplementary Fig. S6B). We hypothesized that the alterations in genes described Moreover, everolimus treatment resulted in increased p-AKT473 in Fig. 3A can be added up to promote aberrant pathway levels, whereas AZD8055 and MLN0128, another mTORC1/2 signaling. Mutations or CNA occurred in PI3K/AKT/mTOR in inhibitor, reduced this upregulation (Supplementary Fig. S6C). 84.9% of tumors, in MAPK pathway in 27.4% of tumors, and AZD8055 and dactolisib, an mTORC1/2-PI3K inhibitor for which in ERBB receptor family of kinases in 42.5% of tumors. OCCC cell lines are highly sensitive (Supplementary Fig. S6A), Combined mutations and CNA in PI3K/AKT/mTOR and MAPK strongly reduced p-AKT473 and to a lesser extent p-AKT308 at pathway genes and the ERBB family of receptor tyrosine high concentrations (Supplementary Fig. S6D). The absence of kinases indicated that 91% of all tumors were affected (Fig. PARP cleavage indicated that everolimus, AZD8055, and

OF6 Clin Cancer Res; 2018 Clinical Cancer Research

Kinome Proﬁling Identiﬁes mTORC1/2 Sensitivity in OCCC

Figure 3. Mutation and CNA distribution. A, Nonsynonymous mutation distribution in genes involved in the frequently mutated PI3K/AKT/mTOR (blue) and MAPK pathway (yellow), the ERBB family of receptor tyrosine kinases (green), and DNA repair pathway (red) as well as ARID1A and PALB2 is shown with OncoPrint. CNA for each mutated gene is added. The 106 OCCC tumors that were both kinome sequenced and SNP arrayed are shown on the horizontal axis ordered on total event frequency in the subsequently altered pathways. (B) The interacting ERBB family of receptor tyrosine kinases, PI3K/AKT/mTOR and MAPK pathways, and (C) DNA repair pathway are commonly altered. These alterations are deﬁned by mutations and CNA.

www.aacrjournals.org Clin Cancer Res; 2018 OF7

Caumanns et al.

dactolisib do not induce apoptosis in an in vitro setting (Supple- mutations were observed in two cell lines, but only OVMANA mentary Fig. S6D). OCCC cell lines presented a large IC50 vari- (R705K) displayed gefitinib sensitivity. In cells with elevated ation for upstream inhibitors of the PI3K/AKT/mTOR and levels of ERBB2 mRNA and high ERBB2 membrane expression MAPK pathway, for example, PI3K inhibitor GDC0941 and MEK (RMG1, SMOV2, and TUOC1), lapatinib sensitivity was observed inhibitor selumetinib, compared with AZD8055 and dactolisib (Fig. 5A). However, some of the tested cell lines without EGFR (Supplementary Fig. S6E). Long-term proliferation inhibition by mutations or elevated ERBB2 levels also displayed sensitivity to dactolisib was stronger compared with AZD8055 (Supplementary gefitinib and lapatinib, indicating involvement of alternative Fig. S6F). Limited sensitivity was observed toward the EGFR mechanisms. The PARP1 inhibitor olaparib showed some efficacy inhibitor gefitinib and ERBB2/EGFR inhibitor lapatinib. EGFR against ATR-mutant TOV21G cells.

Figure 4. OCCC cell line CNA and mutation distribution. A, Chr 1-22 CNA (ampliﬁcations in blue, deletions in red) in OCCC tumors (n ¼ 108) and cell lines (n ¼ 17) depicted from Nexus Copy Number. B, OCCC cell line nonsynonymous mutation distribution in genes involved in the frequently mutated PI3K/AKT/mTOR (blue) and MAPK pathway (yellow), ERBB family of receptor tyrosine kinases (green), and DNA repair (red) pathway as well as ARID1A and PALB2 is shown with OncoPrint. CNA for each mutated gene is added. The 17 OCCC cell lines that were both kinome sequenced and SNP arrayed are shown on the horizontal axis ordered on total event frequency in the subsequently altered pathways. JHOC5 and OV207 are ARID1A mutant but retain ARID1A expression. The genes AKT2, PIK3C3, PIK3CD, PIK3C2B, PIK3R2, PIK3C2G, PIK3C2A, PIK3CG, AKT3, PIK3R4, ERBB4, PALB2, and CHEK1 were not sequenced in JHOC5, HAC2, and OVCA429, and TP53 mutation in TOV21G was detected just above threshold.

OF8 Clin Cancer Res; 2018 Clinical Cancer Research

Kinome Proﬁling Identiﬁes mTORC1/2 Sensitivity in OCCC

Figure 5. OCCC cell line panel inhibitor screening. A, Schematic representation of mutation status, mRNA level, copy-number gain, membrane receptor expression, and a heatmap of inhibitor IC50 in 17 OCCC cell lines. For each cell line horizontally: mutation status of the genes ARID1A, PIK3CA, EGFR, ATR, and TP53 is indicated in gray, EGFR and ERBB2 mRNA levels relative to GAPDH >2 are indicated in green, copy-number gain is indicated in red, and EGFR and ERBB2 membrane expression relative to average mean ﬂuorescence intensity >2 are indicated in yellow. Cell line IC50 for AZD8055 (mTORC1/2), geﬁtinib (EGFR), lapatinib (ERBB2), and olaparib (PARP1/2) are shown horizontally, and cell lines are vertically ordered on AZD8055 sensitivity. B, IC50 of the mTORC1/2 inhibitor AZD8055 from COSMICs (Cancerxgene.org) drug screening database for all cancer cell lines versus ovarian cancer cell lines, horizontal lines indicate geometric mean. C, AZD8055 and everolimus IC50 determined for 14 OCCC cell lines (ES2, KOC7C, SMOV2, JHOC5, RMG1, OVMANA, HAC2, OV207, OVTOKO, TOV21G, OVAS, OVCA429, TUOC1, and RMG2) by MTT assay. HGS ovarian cancer cell lines PEA1, PEO14, and OVCAR3 were used as resistant controls. Horizontal lines indicate geometric mean of only the OCCC cell lines. Data are derived from n 2 experiments. D, Long-term proliferation assay after exposure to increasing concentrations of AZD8055 and everolimus. Results are representative of three experiments.

In conclusion, OCCC cell lines exhibit exquisite sensitivity to datasets (P ¼ 0.053 and P ¼ 0.0028, respectively; Fig. 6A– mTORC1/2 inhibitors, whereas limited sensitivity was observed C). Subsequently, we tested AZD8055 efﬁcacy in OCCC PDX– toward ERBB receptor tyrosine kinase and DNA repair inhibitors bearing NSG mice. Sequencing of 19 genes with the highest despite the high aberration frequency in those pathways. mutation frequency in patients with OCCC showed mutations in ARID1A (1148 stop-gain), PTEN (H93R), and BRCA1 Targeting of mTORC1/2 is effective in PDX models (V356A and 1533 stop-gain) in PDX.155; in PIK3CA Considering the high susceptibility of OCCC cell lines to (K111E, recurrent in OCCC tumors) and ATM (T1020I) in AZD8055, we validated mTORC1/2 activity in OCCC tumor PDX.180; and none in PDX.247. All mutations were present samples. Immunostaining of the mTORC1/2 downstream tar- in the sequenced F0 (patient tumor), F1, and F2 generation. get p-S6 was performed on two different tissue microarrays (n ¼ GISTIC analysis indicated CNA across the PDX models that 136 and n ¼ 83) with primary tumor material from HGS resemble CNA in patients with OCCC (Supplementary ovarian cancer and OCCC. OCCC tumors were more frequently Fig. S7A). Signiﬁcant growth differences were observed between p-S6 positive compared with HGS ovarian cancer in both AZD8055 (10 mg/kg/day) and vehicle-treated mice in all 3 PDX

www.aacrjournals.org Clin Cancer Res; 2018 OF9

Caumanns et al.

Figure 6. p-S6 expression and mTORC1/2 inhibition in OCCC PDX models. A, Representation of p-S6 expression in an OCCC tumor (left) and HGS ovarian cancer tumor (right) as determined by IHC on tissue microarray (TMA). Percentage of (B) positive p-S6 expression in primary OCCC tumors (positive, n ¼ 10; negative, n ¼ 1) and HGS ovarian cancer tumors (positive, n ¼ 76; negative, n ¼ 49) and (C) positive p-S6 expression in a second TMA of primary OCCC (positive, n ¼ 10; negative, n ¼ 1) and HGS ovarian cancer tumors (positive, n ¼ 30; negative, n ¼ 42), determined by the Fisher exact test, two-tailed. Tumor growth, Ki67 expression, and Cleaved Caspase-3 positivity in (D) PDX.155, (E) PDX.180, and (F) PDX.247 tumors following AZD8055 or vehicle treatment. PDX.155 vehicle mice were treated in F4 generation, and AZD8055 mice were treated in F5 generation. Tumor volume is represented as a percentage of initial tumor volume at the start of treatment. Quantiﬁcation of Ki67 and Cleaved Caspase-3 positivity in tumors from vehicle- and AZD8055-treated mice, sacriﬁced on day 21, is shown below respective PDX model tumor growth charts. , P < 0.05; , P < 0.01; and , P < 0.001.

OF10 Clin Cancer Res; 2018 Clinical Cancer Research

Kinome Proﬁling Identiﬁes mTORC1/2 Sensitivity in OCCC

models (Fig. 6D–F). AZD8055 treatment reduced tumor variants in these low-mutation-frequency genes across multiple growth in PDX.247 and inhibited tumor growth in PDX.155 cancer types using the COSMIC database enabled us to predict and PDX.180, which is reflected in a reduction of proliferation AKT1, PIK3R1, FBXW7, ERBB3, ATM, CHEK2,andMYO3A as marker Ki67 positivity (Fig. 6D–F). Remarkably, an increased novel drivers of OCCC (27). Functional effects of AKT1 and staining of the apoptosis marker Cleaved Caspase-3 was PIK3R1 mutations in these domains were described in previous observed in PDX.155 after AZD8055 treatment (Fig. 6D–F). research. For instance, mutations in the AKT1 PH domain were Phosphorylation of mTORC1/2 downstream target S6 was shown to constitutively activate AKT1 (18). Likewise, muta- clearly detectable in the vehicle-treated tumors but did not tions in the PIK3R1 inter-SH1–SH2 domain resulted in loss of significantly change after AZD8055 treatment (Supplementary PIK3R1 function, thus generating PIK3CA hyperactivation (19). Fig. S7B–S7D). These studies support the assumption that AKT1 and PIK3R1 are oncogenic drivers in OCCC. Furthermore, AKT1, PIK3R1, FBXW7,andERBB3 all contribute to mTOR signaling, whereas Discussion ATM, CHEK2,andFBXW7 are implicated in DNA repair sig- Patients with advanced-stage OCCC respond poorly to stan- naling (19, 21, 28). MYO3A mutations are found infrequently dard platinum-based chemotherapy, indicating an unmet need in cancer, but are still implicated in resistance to trastuzumab in for novel treatment strategies. Kinome sequencing of 518 kinases breast cancer (29). and 79 cancer-related genes and SNP array analysis of 108 OCCC In the present study, kinome sequencing and SNP array tumors revealed mutations, amplifications, and deletions in well- analysis revealed substantial percentages of EGFR and ERBB2 known genes such as ARID1A, PIK3CA, TP53, and KRAS and mutations or amplifications in OCCC tumors. EGFR amplifica- several newly identified genes. Collectively, alterations in genes tions (associated with endometriosis, a precursor of OCCC) from the PI3K/AKT/mTOR pathway, the MAPK pathway, and the and ERBB2 amplifications were previously demonstrated in ERBB family of receptor tyrosine kinases were found in 91% of OCCC by comparative genomic hybridization (30, 31). Fur- tumors. In line with these alterations, high mTOR signaling was thermore, sensitivity toward the clinically available inhibitors frequently observed in OCCC, and mTORC1/2 inhibition was gefitinib and lapatinib was shown in OCCC cell lines harboring highly effective in OCCC cell lines and PDX models. These results EGFR or ERBB2 alterations and may be explored further in the indicate that especially targeting mTOR1/2 could be an effective clinic (32, 33). In previous studies, EGFR-targeting therapy such therapeutic approach for patients with OCCC. as gefitinib or erlotinib and EGFR-ERBB2–targeting therapy ARID1A was the most common mutated gene in our OCCC using lapatinib did not provide clinical benefit in pretreated dataset, with a mutation rate of 46.6%. Overall, the mutation patients with ovarian cancer (34–36), but these studies did not rates observed for ARID1A, PIK3CA, TP53,andKRAS matched focus on patients with OCCC. In addition, alterations in DNA earlier sequencing studies on smaller OCCC datasets (3, 4, 6, repair pathway genes were identified in 82% of tumors, sug- 8, 9, 24). Moreover, mutation incidences in PIK3CA, TP53,and gesting opportunities to target DNA repair cascades in OCCC in KRAS correlated with the frequencies found in "pure OCCC" future research. Unfortunately, PARP1/2 inhibition by olaparib tumors as described by Friedlander and colleagues (22), indi- was not effective in our OCCC cell line panel, including BRCA1- cating that our sample cohort also contains pure OCCC tumors. mutant cell lines. Interestingly, inhibition of PARP1/2 by the ARID1A and PIK3CA mutations have been shown as early-onset PARP-trapping agent talazoparib demonstrated efficacy in a mutations in the development of OCCC (23, 25). Previous subset of OCCC cell lines (37). studies have shown that onset of oncogenesis regularly requires Mutations in mTOR have been detected in several tumor types two or more mutational hits in a proto-oncogene or tumor- including HGS ovarian cancer (38). Although mTOR mutations suppressor gene, as was demonstrated in a de novo OCCC were not found in OCCC tumors, alterations in upstream path- mouse model (26). Indeed, in our tumor cohort, 94% of the ways that promote mTORC1/2 activation were found in 91% of PIK3CA- and ARID1A-mutant tumors contained at least one these tumors. In line with these results, abundant high p-S6 additional mutation or CNA in PI3K/AKT/mTOR, MAPK, or expression was observed in OCCC tumors compared with HGS DNA repair pathway genes or the ERBB family of receptor ovarian cancer tumors. Moreover, OCCC cell lines demonstrated tyrosine kinases, which was stage independent. Considering nanomolar range sensitivity toward the mTORC1/2 inhibitor earlier reports and our observed OCCC mutation pattern, we AZD8055 and higher sensitivity compared with HGS ovarian hypothesize that a mutation in ARID1A or PIK3CA is accom- cancer cell lines. To our knowledge, we were the first to exploit panied by either a CNA or mutation in one of the aforemen- the use of PDX models of OCCC. In vivo administration of tioned pathways to promote onset of OCCC. A limitation of the AZD8055 resulted in a significant growth-inhibitory effect in present study can be the imbalance in tumor samples and three OCCC PDX models, consisting of a PIK3CA mutant, an matched controls. Accordingly, we could designate a restricted ARID1A mutant, and a PIK3CA and ARID1A wild-type model. number of mutations as somatic.Notallclinicaldatawere These models reflect the genetic makeup of OCCC. Hence, we available; therefore, we may have underestimated the relation propose mTORC1/2 inhibition as a future treatment strategy in between kinome alteration status and clinicopathologic char- OCCC that should be explored with urgency. Additional inhibi- acteristics or clinical outcome. tion of ARID1A synthetic lethal targets in combination with Genes mutated at low frequency (1%–10%) are generally mTORC1/2 inhibition can be utilized in ARID1A-mutant OCCC difficult to designate as oncogenic drivers. Therefore, we used a tumors (14, 15). large OCCC dataset and binomial testing of observed kinome In the clinic, inhibition of mTORC1 with temsirolimus added variants. Eleven genes were found to be significantly mutated to standard chemotherapy did not improve overall survival in relative to the background mutation rate, and some of these newly diagnosed patients with stage 3 to 4 OCCC as compared had not previously been described in OCCC. Validation of with historical controls (39). However, by blocking both

www.aacrjournals.org Clin Cancer Res; 2018 OF11

Caumanns et al.

mTORC1 and mTORC2, reactivation of PI3K and MAPK signaling Authors' Contributions via mTORC2 can be prevented in OCCC. This may contribute to a Conception and design: J.J. Caumanns, K. Berns, G.B.A. Wisman, M.J. Birrer, better response in patients with OCCC. AZD8055 and another I. Vergote, R. Bernards, A.G.J. van der Zee, S. de Jong mTORC1/2 inhibitor, OSI-027, were tested in non-OCCC cancer Development of methodology: J.J. Caumanns, R.S.N. Fehrmann, H. Itamochi, M.J. Birrer, S. de Jong types. Disappointingly, antitumor efficacy was only observed Acquisition of data (provided animals, acquired and managed patients, above MTD, and both drugs are no longer in clinical development provided facilities, etc.): J.J. Caumanns, K. Berns, H. Klip, E.M. Hijmans, (40, 41). Based on our results, it is conceivable that tumor A.M.C. Gennissen, D. Weening, H.B. Salvesen, I. Vergote, E. van Nieuwenhuy- responses in patients with OCCC can be attained at or below sen, J. Brenton, E.I. Braicu, J. Kupryjanczyk, B. Spiewankiewicz MTD with AZD8055 or OSI-027. A new generation of mTORC1/2 Analysis and interpretation of data (e.g., statistical analysis, biostatistics, inhibitors, such as MLN0128 (Sapanisertib), is now being eval- computational analysis): J.J. Caumanns, G.B.A. Wisman, R.S.N. Fehrmann, T. Tomar, E.M. Hijmans, A.M.C. Gennissen, E.W. Duiker, R.J.C. Kluin, uated in multiple phase II clinical trials (NCT02724020 and fi A.K.L. Reyners, I. Vergote, L. Mittempergher, R. Bernards, S. de Jong NCT02725268). Given the large variation in mutations in speci c Writing, review, and/or revision of the manuscript: J.J. Caumanns, K. Berns, genes of the PI3K/AKT/mTOR and MAPK pathway between G.B.A. Wisman, R.S.N. Fehrmann, E.W. Duiker, A.K.L. Reyners, M.J. Birrer, patients and the high variation in sensitivity toward PI3K and I. Vergote, E. van Nieuwenhuysen, J. Brenton, E.I. Braicu, J. Kupryjanczyk, MEK inhibition in our OCCC cell lines, PI3K or MEK mono- A.G.J. van der Zee, S. de Jong therapy tumor responses in patients with OCCC will presumably Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.J. Caumanns, H. Klip, G.J. Meersma, be limited. MLN0128 is now being combined with a PI3Ka E.M. Hijmans, A.M.C. Gennissen inhibitor (MLN1117) in an ongoing phase II trial in advanced Study supervision: G.B.A. Wisman, R. Bernards, A.G.J. van der Zee, S. de Jong endometrial cancer (NCT02725268). Drugs targeting PI3K as well as mTORC1/2, for example, dactolisib, DS-7423, and XL765, Acknowledgments showed high toxicity in the clinical setting, despite promising We acknowledge Drs. Els Berns, Yasushi Saga, and Vijayalakshmi Shridhar for efficacy in in vitro cancer models, including OCCC (16, 42–45). their kind contribution of OCCC cell lines. We acknowledge the University of Based on our mutational analysis in patients with OCCC, drug Cambridge, National Institute for Health Research Cambridge Biomedical sensitivity screens, and molecular analyses in OCCC cell lines, a Research Centre, Cambridge Experimental Cancer Medicine Centre, and Hutch- ison Whampoa Limited for their support. The authors thank Jolanda A.L. Visser, strategy that combines inhibition of mTORC1/2 with PI3K or Neeltje M. Kooi, Gerda de Vries, and Fernanda X. Rosas Plaza for help with the MEK will have the highest likelihood to improve clinical benefit PDX models. In addition, we acknowledge Anna Piskorz, Mercedes Jimenez- for the majority of patients with OCCC. However, the use of single Linan, and Karen Hosking. G.B.A. Wisman and S. de Jong are members of the target inhibitors of mTORC1/2, PI3K, and MEK at suboptimal EurOPDX Consortium. concentrations is probably crucial to maximize efficacy and This research was supported by grants from the Dutch Cancer Society (KWF, minimize toxicities that were observed with PI3K–mTORC1/2 RUG 2010-4833, 2011-5231, and 2012-5477), the Cock-Hadders foundation, and Cancer Research UK (A15601). dual inhibitors. Overall, our findings set the stage to explore mTORC1/2-targeting phase II clinical trials in OCCC in the future. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in Disclosure of Potential Conflicts of Interest accordance with 18 U.S.C. Section 1734 solely to indicate this fact. E.I. Braicu is a consultant/advisory board member for Clovis, Roche Pharma, and Takeda. No potential conflicts of interest were disclosed by the other Received November 27, 2017; revised March 23, 2018; accepted April 17, authors. 2018; published first April 23, 2018.

References 1. Cancer Genome Atlas Research Network. Integrated genomic analyses of 9. Ho ES, Lai CR, Hsieh YT, Chen JT, Lin AJ, Hung MH, et al. P53 mutation is ovarian carcinoma. Nature 2011;474:609–15. infrequent in clear cell carcinoma of the ovary. Gynecol Oncol 2001; 2. Anglesio MS, Carey MS, Kobel M, Mackay H, Huntsman DG, Speakers 80:189–93. Vancouver Ovarian Clear Cell Symposium. Clear cell carcinoma of the 10. Kuo KT, Mao TL, Chen X, Feng Y, Nakayama K, Wang Y, et al. DNA copy ovary: a report from the first Ovarian Clear Cell Symposium, June 24th, numbers profiles in affinity-purified ovarian clear cell carcinoma. Clin 2010. Gynecol Oncol 2011;121:407–15. Cancer Res 2010;16:1997–2008. 3. Wiegand KC, Shah SP, Al-Agha OM, Zhao Y, Tse K, Zeng T, et al. ARID1A 11. Rahman MT, Nakayama K, Rahman M, Katagiri H, Katagiri A, Ishibashi T, mutations in endometriosis-associated ovarian carcinomas. N Engl J Med et al. Gene amplification of ZNF217 located at chr20q13.2 is associated 2010;363:1532–43. with lymph node metastasis in ovarian clear cell carcinoma. Anticancer Res 4. Jones S, Wang TL, Shih I, Mao TL, Nakayama K, Roden R, et al. Frequent 2012;32:3091–5. mutations of chromatin remodeling gene ARID1A in ovarian clear cell 12. Yamashita Y, Akatsuka S, Shinjo K, Yatabe Y, Kobayashi H, Seko H, et al. carcinoma. Science 2010;330:228–31. Met is the most frequently amplified gene in endometriosis-associated 5. Guan B, Wang TL, Shih I. ARID1A, a factor that promotes formation of SWI/ ovarian clear cell adenocarcinoma and correlates with worsened prognosis. SNF-mediated chromatin remodeling, is a tumor suppressor in gynecologic PLoS One 2013;8:e57724. cancers. Cancer Res 2011;71:6718–27. 13. Helming KC, Wang X, Wilson BG, Vazquez F, Haswell JR, Manchester HE, 6. Kuo KT, Mao TL, Jones S, Veras E, Ayhan A, Wang TL, et al. Frequent et al. ARID1B is a specific vulnerability in ARID1A-mutant cancers. Nat Med activating mutations of PIK3CA in ovarian clear cell carcinoma. Am J Pathol 2014;20:251–4. 2009;174:1597–601. 14. Bitler BG, Aird KM, Garipov A, Li H, Amatangelo M, Kossenkov AV, et al. 7. Hashiguchi Y, Tsuda H, Inoue T, Berkowitz RS, Mok SC. PTEN expression in Synthetic lethality by targeting EZH2 methyltransferase activity in ARID1A- clear cell adenocarcinoma of the ovary. Gynecol Oncol 2006;101:71–5. mutated cancers. Nat Med 2015;21:231–8. 8. Zannoni GF, Improta G, Chiarello G, Pettinato A, Petrillo M, Scollo P, et al. 15.BitlerBG,WuS,ParkPH,HaiY,AirdKM,WangY,etal.ARID1A- Mutational status of KRAS, NRAS, and BRAF in primary clear cell ovarian mutated ovarian cancers depend on HDAC6 activity. Nat Cell Biol carcinoma. Virchows Arch 2014;465:193–8. 2017;19:962–73.

OF12 Clin Cancer Res; 2018 Clinical Cancer Research

Kinome Proﬁling Identiﬁes mTORC1/2 Sensitivity in OCCC

16. Oishi T, Itamochi H, Kudoh A, Nonaka M, Kato M, Nishimura M, et al. The 33. Ryan Q, Ibrahim A, Cohen MH, Johnson J, Ko CW, Sridhara R, et al. FDA PI3K/mTOR dual inhibitor NVP-BEZ235 reduces the growth of ovarian drug approval summary: lapatinib in combination with capecitabine for clear cell carcinoma. Oncol Rep 2014;32:553–8. previously treated metastatic breast cancer that overexpresses HER-2. 17. Sasano T, Mabuchi S, Kuroda H, Kawano M, Matsumoto Y, Takahashi R, Oncologist 2008;13:1114–9. et al. Preclinical efficacy for AKT targeting in clear cell carcinoma of the 34. Posadas EM, Liel MS, Kwitkowski V, Minasian L, Godwin AK, Hussain ovary. Mol Cancer Res 2015;13:795–806. MM, et al. A phase II and pharmacodynamic study of gefitinib in 18. Parikh C, Janakiraman V, Wu WI, Foo CK, Kljavin NM, Chaudhuri S, et al. patients with refractory or recurrent epithelial ovarian cancer. Cancer Disruption of PH-kinase domain interactions leads to oncogenic activation 2007;109:1323–30. of AKT in human cancers. Proc Natl Acad Sci U S A 2012;109:19368–73. 35. Blank SV, Christos P, Curtin JP, Goldman N, Runowicz CD, Sparano JA, 19. Thorpe LM, Yuzugullu H, Zhao JJ. PI3K in cancer: divergent roles of et al. Erlotinib added to carboplatin and paclitaxel as first-line treatment of isoforms, modes of activation and therapeutic targeting. Nat Rev Cancer ovarian cancer: a phase II study based on surgical reassessment. Gynecol 2015;15:7–24. Oncol 2010;119:451–6. 20. Jaiswal BS, Kljavin NM, Stawiski EW, Chan E, Parikh C, Durinck S, et al. 36. Garcia AA, Sill MW, Lankes HA, Godwin AK, Mannel RS, Armstrong DK, Oncogenic ERBB3 mutations in human cancers. Cancer Cell 2013; et al. A phase II evaluation of lapatinib in the treatment of persistent or 23:603–17. recurrent epithelial ovarian or primary peritoneal carcinoma: a gynecologic 21. Cheng Y, Li G. Role of the ubiquitin ligase Fbw7 in cancer progression. oncology group study. Gynecol Oncol 2012;124:569–74. Cancer Metastasis Rev 2012;31:75–87. 37. Wilkerson PM, Dedes KJ, Samartzis EP, Dedes I, Lambros MB, 22. Friedlander ML, Russell K, Millis S, Gatalica Z, Bender R, Voss A. Molecular Natrajan R, et al. Preclinical evaluation of the PARP inhibitor profiling of clear cell ovarian cancers: identifying potential treatment BMN-673 for the treatment of ovarian clear cell cancer. Oncotarget targets for clinical trials. Int J Gynecol Cancer 2016;26:648–54. 2017;8:6057–66. 23. Yamamoto S, Tsuda H, Takano M, Tamai S, Matsubara O. Loss of ARID1A 38. Grabiner BC, Nardi V, Birsoy K, Possemato R, Shen K, Sinha S, et al. A protein expression occurs as an early event in ovarian clear-cell carcinoma diverse array of cancer-associated MTOR mutations are hyperactivating and development and frequently coexists with PIK3CA mutations. Mod Pathol can predict rapamycin sensitivity. Cancer Discov 2014;4:554–63. 2012;25:615–24. 39. Farley JH, Brady WE, Fujiwara K, Nomura H, Yunokawa M, Tokunaga H, 24. Wang YK, Bashashati A, Anglesio MS, Cochrane DR, Grewal DS, Ha G, et al. et al. A phase II evaluation of temsirolimus in combination with carbo- Genomic consequences of aberrant DNA repair mechanisms stratify ovar- platin and paclitaxel followed by temsirolimus consolidation as first-line ian cancer histotypes. Nat Genet 2017;49:856–65. therapy in the treatment of stage III-IV clear cell carcinoma of the ovary. 25. Yamamoto S, Tsuda H, Takano M, Iwaya K, Tamai S, Matsubara O. PIK3CA JCO 2016;34:5531. mutation is an early event in the development of endometriosis-associated 40. Naing A, Aghajanian C, Raymond E, Olmos D, Schwartz G, Oelmann E, ovarian clear cell adenocarcinoma. J Pathol 2011;225:189–94. et al. Safety, tolerability, pharmacokinetics and pharmacodynamics of 26. Chandler RL, Damrauer JS, Raab JR, Schisler JC, Wilkerson MD, Didion JP, AZD8055 in advanced solid tumours and lymphoma. Br J Cancer et al. Coexistent ARID1A-PIK3CA mutations promote ovarian clear-cell 2012;107:1093–9. tumorigenesis through pro-tumorigenic inflammatory cytokine signalling. 41. Mateo J, Olmos D, Dumez H, Poondru S, Samberg NL, Barr S, et al. A Nat Commun 2015;6:6118. first in man, dose-finding study of the mTORC1/mTORC2 inhibitor 27. Fleuren ED, Zhang L, Wu J, Daly RJ. The kinome 'at large` in cancer. Nat Rev OSI-027 in patients with advanced solid malignancies. Br J Cancer Cancer 2016;16:83–98. 2016;114:889–96. 28. Zhang Q, Karnak D, Tan M, Lawrence TS, Morgan MA, Sun Y. FBXW7 42. Kashiyama T, Oda K, Ikeda Y, Shiose Y, Hirota Y, Inaba K, et al. Antitumor facilitates nonhomologous end-joining via K63-linked polyubiquitylation activity and induction of TP53-dependent apoptosis toward ovarian clear of XRCC4. Mol Cell 2016;61:419–33. cell adenocarcinoma by the dual PI3K/mTOR inhibitor DS-7423. PLoS 29. Lapin V, Shirdel EA, Wei X, Mason JM, Jurisica I, Mak TW. Kinome-wide One 2014;9:e87220. screening of HER2þ breast cancer cells for molecules that mediate cell 43. Yu P, Laird AD, Du X, Wu J, Won KA, Yamaguchi K, et al. Characterization of proliferation or sensitize cells to trastuzumab therapy. Oncogenesis the activity of the PI3K/mTOR inhibitor XL765 (SAR245409) in tumor 2014;3:e133. models with diverse genetic alterations affecting the PI3K pathway. Mol 30. Okamoto A, Sehouli J, Yanaihara N, Hirata Y, Braicu I, Kim BG, et al. Cancer Ther 2014;13:1078–91. Somatic copy number alterations associated with Japanese or endometri- 44. Fazio N, Buzzoni R, Baudin E, Antonuzzo L, Hubner RA, Lahner H, et al. osis in ovarian clear cell adenocarcinoma. PLoS One 2015;10:e0116977. A phase II study of BEZ235 in patients with everolimus-resistant, 31. Tan DS, Iravani M, McCluggage WG, Lambros MB, Milanezi F, Mackay A, advanced pancreatic neuroendocrine tumours. Anticancer Res 2016; et al. Genomic analysis reveals the molecular heterogeneity of ovarian clear 36:713–9. cell carcinomas. Clin Cancer Res 2011;17:1521–34. 45. Mehnert JM, Edelman G, Stein M, Camisa H, Lager J, Dedieu JF, et al. A 32. Cohen MH, Williams GA, Sridhara R, Chen G, McGuinn WD, Morse D Jr, phase I dose-escalation study of the safety and pharmacokinetics of a tablet et al. United States Food and Drug Administration Drug approval sum- formulation of voxtalisib, a phosphoinositide 3-kinase inhibitor, in mary: gefitinib (ZD1839; Iressa) tablets. Clin Cancer Res 2004;10:1212–8. patients with solid tumors. Invest New Drugs 2018;36:36–44.

www.aacrjournals.org Clin Cancer Res; 2018 OF13

Integrative Kinome Profiling Identifies mTORC1/2 Inhibition as Treatment Strategy in Ovarian Clear Cell Carcinoma

Joseph J. Caumanns, Katrien Berns, G. Bea A. Wisman, et al.

Clin Cancer Res Published OnlineFirst April 23, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-17-3060

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2018/05/30/1078-0432.CCR-17-3060.DC2

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2018/07/16/1078-0432.CCR-17-3060. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.